README.md

metadata

configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
dataset_info:
  features:
    - name: text_a_embedding
      sequence: float32
    - name: text_b_embedding
      sequence: float32
    - name: prompt_embedding
      sequence: float32
    - name: text_a
      dtype: string
    - name: text_b
      dtype: string
    - name: prompt
      dtype: string
    - name: label
      dtype: int64
  splits:
    - name: train
      num_bytes: 369715678
      num_examples: 39367
    - name: test
      num_bytes: 92475951
      num_examples: 9842
  download_size: 330703108
  dataset_size: 462191629

Glue-STSB with Contrastive Axes

Dataset format:

A pair of sentences, and a prompt along which the sentences are similar or different.

Includes embeddings generated by sentence-transformers.

text_a and text_b are from the Glue-STSB dataset, prompt and label are machine generated.